Mercury Release and Speciation in Chemical ... - ACS Publications

Article pubs.acs.org/EF

Mercury Release and Speciation in Chemical Looping Combustion of Coal T. Mendiara,* M. T. Izquierdo, A. Abad, P. Gayán, F. García-Labiano, L. F. de Diego, and J. Adánez Department of Energy and Environment, Instituto de Carboquímica-ICB-CSIC Miguel Luesma Castán 4, 50018 Zaragoza, Spain ABSTRACT: In the in situ Gasification Chemical Looping Combustion of coal (iG-CLC), the fuel is gasified in situ in the fuel reactor and gasification products are converted to CO2 and H2O by reaction with the oxygen carrier. This work is the first study on mercury release in Chemical Looping Combustion of coal. The determination of mercury speciation (Hg0/Hg2+) in fuel and air reactors as well as the quantification of the amount of mercury released in each reactor were addressed for the first time. The fraction of the mercury in coal vaporized in the fuel reactor depended mainly on the fuel reactor temperature and the coal type. In the fuel reactor, mercury was mainly emitted as Hg0 in the gas phase and the amount increased with the temperature. In the air reactor, mercury was mostly emitted as Hg2+. In a real CLC system, mercury emissions to the atmosphere will decrease compared to conventional combustion as only mercury released in the air reactor will reach the atmosphere. However, measures should be taken to reduce Hg0 in the CO2 stream before the purification and compression steps in order to avoid operational problems.

1. INTRODUCTION Among the Carbon Capture technologies, Chemical Looping Combustion (CLC) has recently emerged as a promising option to facilitate CO2 inherent separation at low cost.1 The use of coal in CLC with carbon capture has been investigated as coal is regarded as increasingly important fuel in the coming years. CLC of coal can be carried out under different configurations. One of them is the in situ Gasification Chemical Looping Combustion (iG-CLC) where two interconnected fluidized bed reactors are used, namely the fuel and air reactors, and an oxygen carrier material transports oxygen from air to fuel reactor. In iG-CLC, coal is fed directly into the fuel reactor and mixed with the oxygen carrier while steam is supplied as fluidizing agent (Figure 1). Alternatively, recirculated CO2 could be also used as fluidizing agent. These gases also act as coal gasifying agents according to the following reaction scheme: coal → volatiles + char

(1)

char + H 2O → H 2 + CO

(2)

char + CO2 → 2CO

(3)

In the iG-CLC process, the fuel combustion is used to produce heat that can be extracted from the CLC reactors and from both flue gases exiting the air and fuel reactors, as these gaseous streams are at high temperature.2 Note that the thermal energy released in a CLC process is the same as in conventional combustion. In the iG-CLC process, unconverted char can be transported to the air reactor mixed with oxygen carrier particles. The release of CO2 in the air reactor as a consequence of the combustion of unconverted char particles negatively affects the carbon capture efficiency of the process. In order to avoid ungasified char reaching the air reactor, a carbon separation unit could be placed after the fuel reactor in an industrial CLC facility.3 In this unit, also known as carbon stripper, char particles are separated from oxygen carrier particles through their differences in fluid dynamic properties. Char particles are then returned back to the fuel reactor while the oxygen carrier particles are transferred to the air reactor for regeneration. In fact, the carbon separation unit is one of the best options to maximize the carbon capture in an iG-CLC system.4 The selection of an adequate oxygen carrier is always a key issue in CLC, but in the case of the iG-CLC process, it becomes even more relevant. Ashes should be drained to avoid their accumulation in the system and some oxygen carrier can be lost together with them. Fe-based oxygen carriers are considered an attractive option due to its low cost and environmental compatibility. Besides, Fe-based materials with magnetic properties could be separated from the ashes. Thus, natural minerals5,6 and industrial wastes5,7 based on iron have been tested as oxygen carriers with different coals showing good performance. In our research group, it was recently reported the use of the Fe-enriched sand fraction from bauxite

The oxygen carrier reacts with the gases produced in reactions 1−3 to yield CO2 and water as products. H 2 , CO, volatiles + n MxOy → CO2 + H 2O + n MxOy − 1 (4)

Once steam is condensed, the gaseous outlet stream from the fuel reactor is a highly concentrated CO2 stream ready for capture. After reaction in the fuel reactor, the reduced oxygen carrier reaches the air reactor, where it is reoxidized again with air. MxOy − 1 +

1 O2 → MxOy 2 © 2014 American Chemical Society

Received: December 3, 2013 Revised: March 6, 2014 Published: March 6, 2014

(5) 2786

dx.doi.org/10.1021/ef402377m | Energy Fuels 2014, 28, 2786−2794

Energy & Fuels

Article

Figure 1. Reaction scheme of the iG-CLC process. Streams where mercury measurements were performed are indicated. OC, oxygen carrier; C, char; dashed line, optional stream.

Sliger et al.19 established that the reaction sequence for Hg oxidation is a Cl-atom recycle process. A detailed thermodynamic study performed by Frandsen et al.17 in conditions similar to those in conventional combustion (including sulfur compounds) established that chlorine enhances vaporization of mercury. Therefore, the chlorine content in coal becomes especially important. High levels of mercury oxidation are most strongly correlated with chlorine concentrations in the coal.20 The averaged chlorine contents in bituminous and lignite coals have been reported as 340 ± 40 and 120 ± 20 ppm, respectively.21 Once a pool of Cl-atoms is established, Hg0 is first oxidized via Cl into HgCl (eq 6) which, in turn, is subsequently oxidized by Cl2 into HgCl2 (eq 7) with an associated regeneration of Cl-atoms. The concentration of atomic Cl was influenced by reactions related to moist CO oxidation, chlorine species transformations and NO formation.22

processing (Fe-ESF) with promising results compared to other materials.4,7,8 Despite their interest, emissions of contaminants in CLC using coal have received little coverage in literature up to date and have been mainly focused on sulfur and nitrogen emissions.9−12 However, the emission of mercury from combustion systems is a topic of growing concern due to the volatility, extremely toxic effects and tendency of mercury to bioaccumulate through the food chain. The accumulation of mercury in the human body can affect human health and generate long-lasting effects. Moreover, mercury emissions are expected to grow in the next future due to the intensive use of coal as fuel for electricity generation.13 Mercury can be present in coal in three forms, namely clays, organic matter, and sulfides (mainly pyrite, FeS2, and cinnabar, HgS).14 Mercury content in coal ranges between 0.02 and 1.0 mg/kg,14 although most of the values are usually found below 0.2 mg/kg. In bituminous coals, mercury is generally associated with FeS2 and HgS, while in sub-bituminous coals mercury is largely associated with the organic fraction.15 These compounds decompose during combustion and mercury is liberated as Hg0 (g). Therefore, the mode of occurrence of mercury does not affect the initial combustion transformation mechanism.16 Nevertheless, elemental mercury Hg0 can be either oxidized or adsorbed on other particles as flue gases cool down. The final speciation of mercury in flue gases may depend on different factors, such as concentration of Hg in the coal, the flue gas temperature and composition, the concentration and physical characteristics of the entrained ash and the time of mercury compounds in the flue gas.16 The release of mercury in both coal gasification and combustion has been studied during recent years. Following thermodynamic calculations, Hg0(g) should be the stable form in the highest temperature regions of combustors and gasifiers.17 In the reducing conditions of a gasification flue gas, Hg0(g) will remain as the dominant form. However, when combustion flue gas cools down, Hg0(g) will react to form Hg22+ and Hg2+ compounds. Nevertheless, the principal oxidized form of mercury is assumed to be Hg2+.16,18 Mercury chemistry involves both homogeneous and heterogeneous reactions. Mercury chlorination has been identified as the main mercury transformation mechanism. Other possible mechanisms are related to mercury interactions with ash particle surfaces where reactive chemical species, catalysts and active sorption sites are available to convert Hg0(g) to Hg2+(g) and particulate mercury, Hg(p).18 According to this, mercury species can be classified into three main forms: Hg0(g), Hg2+(g), and Hg(p).

Hg 0 + Cl → HgCl

(6)

HgCl + Cl 2 → HgCl2 + Cl

(7)

In addition to chlorine, O2 and NO2 can oxidize mercury in the flue gas. The homogeneous reaction between Hg0 and O2 or NO2 was reported to be nonsignificant.18 Nevertheless, these gases facilitate Hg0 adsorption and oxidation in the presence of inorganic and carbonaceous ash particles. The heterogeneous reaction between Hg0 and atomic oxygen to form HgO (s) on particle surfaces was proposed to explain mercury adsorption.18 Other compounds present in gasification and combustion flue gases can also influence mercury oxidation. Moisture has been reported to be a strong inhibitor.22 SO2 had an inhibitory effect when high amounts of HCl are present in the flue gases23 while NO was found to either promote or inhibit homogeneous Hg oxidation, depending on its concentration.19,22 Focusing on the iG-CLC technology, the release of mercury could take place both in fuel and air reactors. The conditions of mercury release would be different in each case as the atmospheres in both reactors are quite different. In the fuel reactor, steam or CO2 would be the major component in the gaseous atmosphere as they are used as gasifying agents. In the air reactor, oxygen will be present in high concentration. The mercury found at the fuel reactor outlet will affect the quality of the CO2 to be compressed, transported and stored which can be a concern for the CO2 processing unit. Mercury can accumulate in the CO2 processing unit and attack the aluminum heat exchangers. Moreover, it is known that trace quantities of heavy metals can participate in mineralization and precipitation reactions in sequestration conditions.24 The 2787


Energy & Fuels

Article

were connected by an U-shaped fluidized bed used as loop seal (2). The loop seal avoided mixing the different atmospheres present in fuel and air reactors. The total amount of oxygen carrier introduced to the CLC system was 3.5 kg Fe-ESF, around 1.2 kg in the fuel reactor. The bubbling fluidized bed acting as fuel reactor (FR) was made of steel (50 mm ID and 200 mm bed height). Anthracite was fed to the fuel reactor by a screw feeder just above the distributor plate at the bottom of the bed. In these experiments, 110 g/h of anthracite were fed, which correspond to a thermal power of 550 Wth. Steam was produced in an evaporator (10) and also introduced at the bottom of the fuel reactor as coal gasifying agent. The flow to the fuel reactor corresponded to a velocity of 0.1 m/s at 900 °C. The gaseous products generated in coal gasification reacted with the oxygen carrier to produce CO2 and H2O, while the oxygen carrier was reduced. Then, the oxygen carrier was transferred through the loop seal (50 mm ID) to the air reactor (80 mm ID and 100 mm bed height), also made of steel. In the air reactor, the oxygen carrier was regenerated in air. The air flow introduced corresponded to a velocity of 0.5 m/s at 900 °C. The ungasified char particles that may have been transferred to the air reactor together with the reduced oxygen carrier might be also combusted releasing CO2 in the air reactor. To complete the chemical loop, the regenerated oxygen carrier particles were sent back to the fuel reactor. A secondary air stream was injected at the top of the bubbling fluidized bed acting as air reactor to help particle entrainment. Particles were thus sent up the riser (4) and separated from the gaseous stream in a high-efficiency cyclone (5). The rate of circulation of the oxygen carrier in the unit was measured by a diverting solid valve (6) located after the cyclone. In these experiments, the circulation flow rate of Fe-ESF was maintained at around 4.5 kg/h. The gaseous stream from the air reactor was sent through a filter (12) to the stack. The filter used quartz wool as filter medium and was maintained at 200−400 °C. On the other hand, the particles collected in the cyclone were accumulated in a solid reservoir (7), which also served as seal to avoid mixing atmospheres from air and fuel reactors. The amount of oxygen carrier sent back to the fuel reactor was regulated by a solids control valve (8). As it can be observed in Figure 2, the ICB-CSIC-s1 unit does not include a carbon separation unit after the fuel reactor. Nevertheless, the absence of a carbon separation unit in our CLC unit makes the interpretation of the effect of different operational conditions on the results obtained easier. The temperature in the air reactor was maintained at around 950 °C in all the tests and the fuel reactor temperature was varied from 875 to 930 °C. A total of 32 h of continuous fluidization were registered of which 18 h corresponded to fuel combustion. CO, CO2, H2, CH4, and SO2 concentration in the fuel reactor outlet stream together with CO, CO2, O2, and SO2 from the air reactor were recorded. Gray boxes in Figure 2 indicate the points where mercury measurements were performed in the ICB-CSIC-s1 unit. Sampling of gaseous mercury (Hg0 and Hg2+) in the gaseous outlet stream from fuel and air reactors was performed once steady state conditions were reached at each temperature. Then, the mercury content in the outlet stream of each reactor was measured during 30 min. The sampling was independently carried out, so that sampling one stream does not affect the results of the other stream. As shown in Figure 2, a slip stream from the main outlet streams from fuel and air reactors was diverted to the analyzers for continuous online measurement. Elemental mercury (Hg0) was monitored online by a cold vapor atomic absorption spectroscopy (CVAAS) apparatus, VM3000 from MCI (Hg0_FR and Hg0_AR in Figure 2). Moreover, oxidized mercury in the slip stream was determined according to Ontario Hydro Method. The slip streamflow was measured all the time. Glass impingers containing 1 M KCl were placed in a bath at 0 °C to retain oxidized mercury (Hg2+_FR and Hg2+_AR in Figure 2). Mercury in anthracite (Hg_coal in Figure 2) and in solids elutriated from air reactor (Hg_AR,s in Figure 2) were also determined. Mercury content was directly analyzed on solids and impingers solutions using an automatized mercury analyzer LECO AMA254 gold amalgam atomic absorption spectrometer. All the tubes and fittings in the sampling train were made of Teflon, including the internal connections of the CVAAS apparatus. This fact avoided mercury adsorption on fittings.

mercury released through the air reactor will be emitted to the atmosphere. Many efforts have been conducted during last years in order to regulate mercury emissions from coal-fired power plants. As a result, mercury-specific legislation was already set in United States, Canada, and Germany. In 2012, China adopted the air pollution emission standards for coalfired power plants, including mercury. Therefore, the investigation of the fate of mercury in CLC systems is especially relevant. It would be interesting to know both the amounts and the speciation, Hg0(g) and Hg2+(g), in both fuel and air reactors. The present work studies for the first time the fate of mercury in coal when it is burned in an iG-CLC system and analyzes the differences with respect to conventional combustion. Special attention will be paid to the influence of the fuel reactor temperature on the mercury emissions during the process.

2. MATERIALS AND METHODS 2.1. Oxygen Carrier and Coal Used. The Fe-enriched sand fraction (Fe-ESF) from bauxite processing in the alumina production was supplied by Alcoa Europe-Alúmina Española S.A. This bauxite waste was dried at room temperature for 72 h and then sieved to the desired size (+100−300 μm). The dried sample was calcined at 1200 °C during 18 h to ensure complete oxidation of the sample and to increase the resistance to attrition. The resulting material was used as oxygen carrier in the iG-CLC process and was activated before its use in the present experiments.8 The main chemical and physical properties of the material are shown in Table 1.8 Anthracite from

Table 1. Characterization of the Bauxite Waste (Fe-ESF) Fe2O3 (wt %)a XRD main phases crushing strength (N) oxygen transport capacity, ROC (%)b porosity (%) skeletal density (kg/m3) specific surface area, BET (m2/g) a

58 Fe2O3, β-Al2O3 4.3 2 10 4266 0.1

Determined by TGA. bTransformation to Fe3O4

“El Bierzo” (Spain) and lignite from Teruel (Spain) were used as fuel. They were milled and sieved to the desired size (+200−300 μm). Ultimate and proximate analyses of the coals are shown in Table 2, together with the low heating value. 2.2. Experimental Setup and Procedure. Anthracite combustion experiments were performed at the ICB-CSIC-s1 unit. A scheme of this unit is shown in Figure 2. The chemical looping system is composed of two fluidized beds, fuel (1) and air reactors (3), which

Table 2. Proximate and Ultimate Analysis of the Anthracite and Lignite Used (wt % Except Indicated) elemental (wt %) carbon hydrogen nitrogen sulfur oxygena mercury (mg/kg) chlorine (mg/kg) LHV (kJ/kg) a

anthracite

lignite

60.7 2.0 0.9 1.3 2.7 0.274

45.4 2.5 0.6 5.2 9.9 0.108

183

63

21878

16251

proximate (wt %)

anthracite

lignite

moisture ash volatile matter fixed carbon

1.0 31.5 7.5 59.9

12.5 25.2 28.6 33.6

By difference. 2788


Energy & Fuels

Article

Figure 2. Experimental setup ICB-CSIC-s1. Gray boxes indicate the points where mercury measurements were performed.

3. RESULTS 3.1. Analysis of the Anthracite Combustion. The evaluation of the combustion performance is interesting regarding the interpretation of mercury emissions in both fuel and air reactors, as this evaluation can give an indication of the degree of coal conversion and the presence of product/ unburned compounds that may be generated and influence the mercury chemistry. The experiments with anthracite were carried out at different temperatures in the range 875−930 °C. All the tests were performed at stoichiometric conditions, defined by a value of the oxygen-carrier to fuel ratio (ϕ) equal to unity (see eq 8). This ratio compares the oxygen supplied by the circulating oxygen carrier to the oxygen needed to burn the fuel fed. In this case, it is assumed that the state of reduction of the oxygen carrier in all cases was Fe3O4. ϕ=

FOCR OC ṁ SFΩSF

(8)

Figure 3 shows a typical product distribution both in fuel and air reactors obtained once the steady state was reached in the experiment at 920 °C. As it can be seen, the operation of the experimental unit was smooth and the values remained quite stable during the 30 min measurement. CO2 was the main product in the fuel reactor. The main unburned compounds were CO, H2, and CH4. It is expected that the CO2 molar flow increases with temperature while the amount of unburned compounds decreases.8 SO2 was detected at the outlet of the fuel reactor and reached a value close to 6600 ppm at 920 °C. At the air reactor outlet SO2 was also detected. It was formed in the combustion of the char not gasified that reached the air reactor. Gasification is enhanced at higher temperatures in the fuel reactor and lower SO2 emissions in the air reactor are registered.12 At 920 °C, around 60 ppm were recorded at the

Figure 3. Evolution with time of the gas concentration (dry basis and N2 free) of the main gases in the fuel reactor (CO, CO2, H2, CH4, SO2, and Hg0/Hg2+) and in the air reactor (O2, CO2, SO2, Hg0/Hg2+) at TFR = 920 °C.

air reactor outlet. Both Hg0 and Hg2+ concentrations were also included in Figure 3. In the case of Hg2+ and according to the 2789


Energy & Fuels

Article

oxygen lacking to achieve complete combustion to CO2 and H2O of the gaseous compounds in the fuel reactor and the oxygen demand of the coal introduced.

measurement procedure previously described, the averaged value for the whole measurement period was plotted. The combustion performance of the Fe-ESF oxygen carrier burning anthracite can be analyzed focusing on two parameters: the carbon capture efficiency (ηCC) and the total oxygen demand (ΩT). The carbon capture efficiency, ηCC, is the fraction of the total carbon fed converted to gas in the fuel reactor. Only CH4, CO, and CO2 were considered, as the amount of tars and hydrocarbons heavier than CH4 detected was negligible.8 ηCC =

ΩT =

(9)

The carbon capture depends on the carbon in the char formed, which is already gasified, represented by the char conversion (Xchar). All the volatiles in coal are assumed to be captured. Char conversion is calculated as the fraction of carbon in the char, which is released to the FR exhaust gas stream. Xchar =

(ṁ SFΩSF/MO2)

(11)

The oxygen demand decreased with increasing temperatures. In general, very low values were observed, around 1%. The oxygen demand is mainly attributed to the volatiles that escape from the fuel reactor without being converted. In the case of anthracite, the amount of volatiles is low and the Fe-ESF particles present high reactivity to the main gases in the volatiles,25,26 what would explain this low values of oxygen obtained. 3.2. Analysis of the Mercury Emissions. As it has been already mentioned, the objective of this study is to analyze mercury release in the fuel and air reactors of an iG-CLC system burning coal. The study of the mercury release using anthracite as fuel was carried out at two different temperatures in the fuel reactor (i.e., 875 and 920 °C). The streams where the mercury content was determined were shown on both Figures 1 and 2. First, the mercury content in the solids fed to the system was analyzed. Anthracite presented high mercury content (Hg_coal = 0.274 mg/kg coal). The mercury content in the Fe-ESF was negligible, lower than 0.0001 mg/kg. Finally, the mercury present in the fines at the outlet of the air reactor was also measured. These solids are mainly composed by coal ash. At all the temperatures analyzed, the mercury content in the fines (Hg_AR,s) was under the measure limit of the equipment (lower than 0.0001 mg/kg) and can be considered as negligible. In the gaseous outlet streams, the mercury content and the speciation (Hg0/Hg2+) at the exit of the fuel and air reactors was determined following the experimental procedure indicated in section 2.2. Figure 5A shows the weight percentages of the different mercury species in the exhaust streams from both fuel and air reactors referred to the total mercury in the anthracite fed at the two temperatures tested. Note that in all the experiments performed, mercury balances closed between 80 and 95%. These values are inside the interval proposed in order to consider the balance correct (70−130%).27−29 The standard deviation of the mercury measurements performed using the experimental methodology detailed in this work was calculated. For the calculation, six measurements were performed under similar conditions. The standard deviation was estimated as 6%. The analysis of the results should consider the different atmospheres present in fuel and air reactors. In the fuel reactor, coal gasification with steam takes place in the presence of the oxygen carrier (Fe-ESF). Therefore, mercury release in this reactor would be compared to that expected in a conventional coal gasification unit. According to thermodynamics, it would have been expected that mercury vaporized completely once coal was introduced into the fuel reactor.17 However, the first relevant result that can be extracted from Figure 5A is that not all the mercury in coal was released during coal gasification in the fuel reactor. Only around half of the total amount in average was measured there. The amount of mercury released from anthracite in the fuel reactor increased with temperature. It was 39.5% of the total mercury at 875 °C and 56.6% at 920 °C. This value of mercury release in the fuel reactor at the highest temperature corresponds to a low value of char conversion, around 13% as seen in Figure 4. Another

[FCO2,FR + FCO,FR + FCH4,FR ] [FCO2,FR + FCO,FR + FCH4,FR + FCO2,AR ]

0.5FCO,FR + 2FCH4,FR + 0.5FH2,FR

[FCO2,FR + FCO,FR + FCH4,FR − FC,vol] [FCO2,FR + FCO,FR + FCH4,FR + FCO2,AR − FC,vol] (10)

Figure 4 presents the values of the carbon capture (ηCC) and char conversion (Xchar) for different fuel reactor temperatures in

Figure 4. Carbon capture efficiency (ηCC) ■, char conversion (Xchar) Δ, and total oxygen demand (ΩT) ● in experiments at different FR temperatures. Conditions: ϕ = 1.

the combustion of this anthracite. In general, an increase in temperature leads to an increase in the char conversion and the carbon capture efficiency. Nevertheless, the carbon capture values reached were low, around 20% at the highest temperature tested. As it has been already observed,25 the char gasification rate of this anthracite is low, and therefore, there is a significant part of the char produced in the gasification process that reaches the air reactor. This causes the carbon capture efficiency to decrease. In order to reach higher values, the use of a carbon separation unit would be advisable,3 although the analysis of its effect is not the objective of the present study. It is also relevant to note that the carbon capture and the char conversion values are very close. This is due to the low volatile content of this coal, see Table 2. Figure 4 also presents the values for the total oxygen demand (ΩT). This parameter was calculated as the quotient between the 2790


Energy & Fuels

Article

Figure 5. Total mercury emissions in FR (fuel reactor) and AR (air reactor) at different temperatures referred to the total mercury in the coal fed in the experiments with (A) anthracite and (B) lignite.

920 °C. The chlorine content in anthracite according to Table 2 was 183 mg/kg. The fuel reactor bed consists of Fe-ESF whose major constituents are Fe2O3 and Al2O3. In principle, the presence of this oxygen carrier may contribute to a larger extent of mercury oxidation when compared to conventional gasification, although in this case the surface area of the Fe-ESF material is low (0.1 m2/g). However, the effect of both the increase in temperature and the presence of Fe-ESF seem to be hindered. One of the possible explanations is the presence of SO2 in the flue gases of the fuel reactor. As mentioned before, SO2 can inhibit mercury oxidation if high amounts of chlorine are present in the flue gases.23 This may be the case in these experiments. At 920 °C, around 6600 ppm of SO2 were detected at the outlet of the fuel reactor and the amount of SO2 is expected to increase with temperature, as the transformation of H2S to SO2 is favored by temperature.12 Following this SO2 inhibition tendency, the extent of Hg0 oxidation would decrease with the increase in the fuel reactor temperature, as it was observed in Figure 5A. Following the previous results, Figure 5A shows the mercury emitted in the air reactor at the two fuel reactor temperatures analyzed. In the air reactor, air is introduced to regenerate the oxygen carrier. The char transported to this reactor will be also burned here in the reaction with air. Thus, mercury emissions in the air reactor would be comparable to those obtained in a conventional combustion system such as pulverized coal fired boilers. Measurements performed in several coal combustion pilot- and full-scale systems before the air pollution control devices (APCD) showed that the gaseous and particulate Hg2+ compounds, Hg2+X (s,g), generally represent more than half of the total mercury present in combustion flue gases. The other predominant form of mercury in combustion flue gas is Hg0(g).16 In the present experiments, the temperature in the air reactor was maintained constant at 950 °C. It can be assumed that at this temperature almost all of the mercury was released, as no mercury was found in the fly ash escaping from the air reactor (Hg_AR,s). The amount of mercury released in the air reactor gases decreases with the increase in the fuel reactor temperature as more mercury had been previously released in the fuel reactor. The mercury released in the air reactor represents 40.5% of the total mercury at 875 °C and 38.7% at 920 °C. The mercury speciation found in the air reactor was also presented in Figure 5A. In this case, the molar ratio Hg0/ Hg2+ was 38/62 at the two fuel reactor temperatures. Oxidized

interesting fact that can be obtained from Figure 5A is that the major species in the fuel reactor is Hg0 (g). The percentage of mercury released as Hg0 (g) in the fuel reactor increases with temperature. The molar ratio Hg0/Hg2+ in the gas phase was 76/24 at 875 °C and 86/14 at 920 °C. The volatility of mercury is known to be influenced by several variables such as temperature, coal composition, and atmosphere. Li et al.29 analyzed the effect of these variables on mercury volatility and speciation in a detailed study testing five different Chinese coals commonly used in gasification units. They concluded that the most influencing variable on mercury volatility was temperature. At 800 °C, they reported volatilities between 75 and 85%, depending on the type of coal, while at 1200 °C the volatility was around 95% for all the coals. According to their results, some mercury remained in the char at the different temperatures tested. The mercury content in the coals also affected mercury volatility. The authors reported and increase in volatility with the increasing mercury content in most of the conditions tested. Besides temperature and mercury content, mercury volatility was found to depend on the mode of occurrence of mercury. Mercury in pyrite was more difficult to release than mercury in more hydrophilic minerals.29−31 Moreover, steam may inhibit mercury volatility compared to oxygen.30,32 Other factors may affect mercury volatility, but in much lesser extent. This is the case of the residence time.29 Results presented in Figure 5A agree with some of the findings summarized above. The mercury content of the anthracite is high (Table 2), which would favor mercury volatility. However, a strong affinity of mercury for sulfide association has been reported for anthracites.33 The pyritic sulfur content is around 0.5 wt % in this coal, which may contribute to inhibit mercury volatility. The major presence of Hg0 (g) in the fuel reactor is in accordance with thermodynamics, which points to Hg0 as the main species in gasification flue gases.16,17 It has been observed that in gasification environments, the increase in temperature increased the percentage of oxidized mercury, especially when the chlorine content in coal was high.29 Moreover, although there is some controversy about the effect of the iron species,19,34 iron ore and alumina compounds have been identified as promoters of the conversion of Hg0 to Hg2+X, where X is Cl2 or O.35 However, the effects described above are not observed in the present experiments, as the molar ratio Hg0/Hg2+ in the gas phase was 76/24 at 875 °C and 86/14 at 2791


Energy & Fuels

Article

reactor will be present in the CO2 stream intended to be captured. Directive 2010/75/UE sets the limits for mercury emissions from incineration systems to 0.05 mg/Nm3 (normalized to 10% O2 vd. in the outlet stream). In the absence of specific limits for combustion systems, this value can be taken as a reference. Table 3 gathers the emission values measured in the air reactor

mercury is the major species. The presence of significant excess oxygen would favor mercury catalytic oxidation. Although SO2 was also detected in the air reactor flue gas, the amount was around 100 times less than that found in the fuel reactor, which would also minimize the inhibition effect of this gas on mercury oxidation. In order to further evaluate mercury emissions in CLC systems, the measurements using Fe-ESF and anthracite were compared to those obtained in the same experimental unit using ilmenite as oxygen carrier and Spanish lignite as fuel.12 The mercury content in the lignite used was 0.11 mg/kg and the chlorine content was 63 mg/kg, both less than in the case of anthracite. The mercury content in ilmenite was negligible. The temperature for the experiment performed with lignite was an intermediate temperature between those used with anthracite, that is, 910 °C. The comparison of mercury measurements at the outlet stream of fuel and air reactors using lignite is shown in Figure 5B. Regarding mercury emission in the fuel reactor, the first important result in the experiments with lignite is that only around of 31.5% of the mercury in the coal was emitted, although char conversion reached 87%. This value was lower than 39.5% found in the experiments with anthracite at 875 °C and in this case, char conversion was 7%. According to this, char conversion does not seem to affect mercury release. Temperature would be the key variable influencing mercury volatilization. But according to this, higher mercury release would be expected in the experiments with lignite at 910 °C. The lower value found can be explained considering several factors. First, the larger mercury content of the anthracite compared to lignite and especially, the higher content of pyritic sulfur in the lignite used (1.9 wt % in the coal), which would notably difficult mercury volatilization. The residence time of the coal particles in the fuel reactor was also shorter in the experiments with lignite, but the difference is not significant enough to justify the decrease in mercury volatility. The molar ratio Hg0/Hg2+ in the fuel reactor was calculated and resulted 75/25. It is interesting to highlight that the major mercury species in the fuel reactor atmosphere was again Hg0 (g). Important mercury oxidation inhibition by SO2 would also be expected in the experiments with lignite, as the concentration of this gas in the fuel reactor was very high.12 In the oxidizing environment of the air reactor, Hg2+ was again the most abundant species. The molar ratio Hg0/Hg2+ measured in the air reactor was 46/54. The analysis of mercury release with different coals, such as anthracite and lignite allow to draw some interesting conclusions for the iG-CLC process. From the findings presented above it can be expected that mercury in coal will be split and released both in fuel and air reactors. The split ratio will mainly depend on the type of coal used and foremost on the fuel reactor temperature. The main species that could be found in the fuel reactor would be Hg0(g) while in the air reactor the predominant species would be Hg2+(g).

Table 3. Mercury Concentration in the Fuel (Dry Basis) and Air Reactors in iG-CLC Experiments with Different Fuels coal

TFR (°C)

HgFR (mg/Nm3)

HgAR (mg/Nm3) (10% vd.)

anthracite

875 920 910

0.082 0.084 0.015

0.006 0.005 0.011

lignite

in the different experiments presented in this paper. Mercury emissions in the air reactor are well below the 0.05 mg/Nm3 and would not be a problem in CLC systems, even in the case of using coals with high mercury content such as the anthracite used in this work. In the case that the outlet stream from the air reactor should be treated to control other emissions such as NOx, particulate matter or sulfur emissions, mercury emission could even be reduced. It has been already reported that combustion systems equipped with SCR, ESP, and wet FGD significantly contribute to increase mercury removal efficiency.27 On the other hand, the mercury released in the fuel reactor affects CO2 quality. In the past years, several studies have addressed the analysis of the quantities of other compounds that may be present in a CO2 stream intended for capture and storage. The presence of these compounds can affect energy requirements in transport and storage. De Visser et al.36 presented some recommendations for pipeline CO2 transportation. In these recommendations, no values for mercury were included, although it is already known that the presence of trace elements can cause important operational problems related to liquid metal embrittlement in the aluminum components. There are three different levels where mercury, both elemental and ionic, could be captured in the CO2 processing unit: the compression stage, the moisture removal stage and by activated carbon injection (ACI), which has revealed as one of the most promising options for mercury removal.20 Activated carbon beds have been commonly employed in the natural gas industry to reach concentrations less than 0.01 μg/m3, which is the limit settled to avoid operational problems. Table 3 also includes the total concentration of mercury that was measured in the fuel reactor in the experiments with different fuels. Real CLC systems will include a carbon separation unit, where unconverted char from the fuel reactor would be separated from the oxygen carrier particles and returned back to the fuel reactor. The char would therefore increase the residence time in the fuel reactor. It is expected that the carbon separation unit improves the carbon capture efficiency of the process but regarding mercury emissions, no important effect is foreseen as the increase in the residence time has not a great effect on mercury volatility at a certain temperature. Thus, mercury splitting seems to be unaffected by char conversion. Not for nothing, with lignite the majority of mercury is evolved in the air reactor but char was highly converted in the fuel reactor in this case.12 Nevertheless, the mercury content measured in the fuel reactor outlet (dry basis) was much larger than that demanded in order to avoid

4. MERCURY EMISSIONS AND CO2 QUALITY The results shown are relevant for addressing Hg emission control in the CLC system and also regarding the quality of the CO2 obtained. The identification and quantification of Hg speciation forms done in the present work is very important as Hg0 and Hg2+ have distinctive physical, chemical, and biological properties.16 Mercury emission to the atmosphere may take place in the air reactor while the mercury released in the fuel 2792


Energy & Fuels

Article

FOC = solid circulation rate (kg/h) MO2 = molar mass of molecular oxygen (0.032 kg/mol) ṁ SF = coal mass flow (kg/s) ROC = oxygen transport capacity Xchar = char conversion

corrosion problems in the pipelines. Therefore, it can be anticipated that the CO2 stream from the fuel reactor should be treated to reduce its mercury content. The results obtained in this work are of relevance as they allow us to predict that the release of mercury in a CLC system will be more an operational than an environmental issue. In a CLC system, mercury emissions will be significantly reduced compared to conventional combustion as only the mercury released in the air reactor will be emitted to the atmosphere.

Greek Symbols

ϕ = oxygen-carrier to fuel ratio ΩSF = coal oxygen demand (kg oxygen/kg coal) ΩT = total oxygen demand (%) ηCC = carbon capture efficiency

5. CONCLUSIONS Mercury release in a continuous iG-CLC unit using anthracite and lignite as fuels and low cost Fe-based oxygen carriers was analyzed. It was found that the amount of mercury released in the fuel reactor depended mostly on the fuel reactor temperature and the coal used. Values around 39.5 and 56.6% of the total mercury were found in experiments with anthracite. The increase in temperature resulted in an increase of the amount of mercury released in the fuel reactor. However, a high pyrite content of the coal negatively affected mercury volatility. Mercury speciation was also measured as it is the key factor determining the mercury control measurements to adopt. In the fuel reactor, Hg0 was found to be the major species. In this case, it is speculated that the high SO2 concentration in the experiments with anthracite and lignite may play some inhibition effect on elemental mercury oxidation to Hg2+. In the air reactor, with oxidizing conditions, mercury will be found mainly as Hg2+. Considering the total mercury concentrations at the outlet of both reactors found in this work, no problems with mercury emissions in the air reactor are anticipated if the legislation corresponding to incineration systems is considered as reference. Nevertheless, the mercury content in the fuel reactor outlet stream should be reduced to avoid operational problems in the CO2 processing unit.

■

Acronyms

■

AR = air reactor FR = fuel reactor LS = loop seal OC = oxygen carrier C = char

REFERENCES

(1) Adánez, J.; Abad, A.; García-Labiano, F.; Gayán, P.; de Diego, L. F. Prog. Energy Combust. 2012, 38, 215−282. (2) Abad, A.; Adánez, J.; García-Labiano, F.; de Diego, L. F.; Gayán, P.; Celaya, J. Chem. Eng. Sci. 2007, 62, 533−549. (3) Cao, Y.; Pan, W. P. Energy Fuels 2006, 20, 1836−1844. (4) Abad, A.; Cuadrat, A.; Mendiara, T.; García-Labiano, F.; Gayán, P.; de Diego, L. F.; Adánez, J. Ind. Eng. Chem. Res. 2012, 51, 16230− 16241. (5) Leion, H.; Jerndal, E.; Steenari, B. M.; Hermansson, S.; Israelsson, M.; Jansson, E.; et al. Fuel 2009, 88, 1945−1954. (6) Berguerand, N.; Lyngfelt, A. Fuel 2008, 87, 2713−2726. (7) Mendiara, T.; Pérez, R.; Abad, A.; de Diego, L. F.; GarcíaLabiano, F.; Gayán, P.; Adánez, J. Ind. Eng. Chem. Res. 2012, 51, 16216−16229. (8) Mendiara, T.; de Diego, L. F.; García-Labiano, F.; Gayán, P.; Abad, A.; Adánez, J. Int. J. Greenhouse Gas Control 2013, 17, 170−182. (9) Shen, L.; Gao, Z.; Wu, J.; Xiao, J. Combust. Flame 2010, 157, 853−863. (10) Song, T.; Shen, L.; Xiao, J.; Chen, D.; Gu, H.; Zhang, S. Combust. Flame 2012, 159, 1286−1295. (11) Adánez-Rubio, I.; Abad, A.; Gayán, P.; García-Labiano, F.; de Diego, L. F.; Adánez, J. Appl. Energy 2014, 113, 1855−1862. (12) Mendiara, T.; Izquierdo, M. T.; Abad, A.; de Diego, L. F.; García-Labiano, F.; Gayán, P.; Adánez, J. Int. J. Greenhouse Gas Control 2014, 22, 15−24. (13) IEA. World Energy Outlook (WEO) 2012. Executive summary; IEA: Paris, 2012. (14) Yudovich, Y. E.; Ketris, M. P. Int. J. Coal Geol. 2005, 62, 107− 134. (15) Senior, C.; Eddings, E. Evolution of Mercury from Limestone, PCA R&D Serial No. 2949; Portland Cement Association: Skokie, IL, 2006. (16) Galbreath, K. C.; Zygarlicke, C. J. Environ. Sci. Technol. 1996, 30, 2421−2426. (17) Frandsen, F.; Dam-Johansen, K.; Rasmussen, P. Prog. Energy Combust. Sci. 1994, 20, 115−138. (18) Galbreath, K. C.; Zygarlicke, C. J. Fuel Process. Technol. 2000, 65−66, 289−310. (19) Sliger, R. N.; Kramlich, J. C.; Marinov, N. M. Fuel Process. Technol. 2000, 65−66, 423−438. (20) Pavlish, J. H.; Sondreal, E. A.; Mann, M. D.; Olson, E. S.; Galbreath, K. C.; Laudal, D. L.; et al. Fuel Process. Technol. 2003, 82, 89−165. (21) Yudovich, Y. E.; Ketris, M. P. Int. J. Coal Geol. 2006, 67, 127− 144. (22) Niksa, S.; Helble, J. J.; Fujiwara, N. Environ. Sci. Technol. 2001, 35, 3701−3706. (23) Zhou, J.; Luo, Z.; Hu, C.; Cen, K. Energy Fuels 2007, 21, 491− 495.

AUTHOR INFORMATION

Corresponding Author

*Phone: + 34 976 733 977. Fax: +34 976 733 318. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors thank the Government of Aragón and La Caixa (2012-GA-LC-076 project) and the Spanish Ministry for Science and Innovation (ENE2010-19550 project) for the financial support. T. Mendiara is thankful for the “Juan de la Cierva” postdoctoral contract awarded by this Ministry. P. Gayán thanks CSIC for the financial support of the project 201180E102. The authors also thank to Alcoa Europe-Alúmina Española S.A. for providing the Fe-enriched sand fraction used in this work. G. Galo is acknowledged for his contribution to the experimental results.

■

NOMENCLATURE FCO2,AR = carbon dioxide molar flow in the air reactor (mol/s) FC,vol = carbon flow from the volatile matter (mol/s) Fi,FR = i species molar flow in the fuel reactor outlet stream (mol/s) 2793


Energy & Fuels

Article

(24) Reddy, K. J.; John, S.; Weber, H.; Argyle, M. D.; Bhattacharyya, P.; Taylor, D. T.; et al. Energy Procedia. 2011, 4, 1574−1583. (25) Mendiara, T.; García-Labiano, F.; Gayán, P.; Abad, A.; de Diego, L. F.; Adánez, J. Fuel 2013, 106, 814−826. (26) Mendiara, T.; Abad, A.; de Diego, L. F.; García-Labiano, F.; Gayán, P.; Adánez, J. Energy Fuels 2012, 26, 1420−1431. (27) Lee, S. J.; Seo, Y.-C.; Jang, H.-N.; Park, K.-S.; Baek, J.-I.; An, H.S.; Song, K.-C. Atmos. Environ. 2006, 40, 2215−2224. (28) Zhang, L.; Zhuo, Y.; Chen, L.; Xu, X.; Chen, C. Fuel Process. Technol. 2008, 89, 1033−1040. (29) Li, Y.; Zhang, J.; Zhao, Y.; Zheng, C. Energy Fuels 2011, 25, 3988−3996. (30) Sekine, Y.; Sakajiri, K.; Kikuchi, E.; Matsukata, M. Powder Technol. 2008, 180, 210−215. (31) Bunt, J. R.; Waanders, F. B. Fuel Process. Technol. 2011, 92, 1646−1655. (32) Helble, J.; Mojtahedi, W.; Lyyränen, J.; Jokiniemi, J.; Kauppinen, E. Fuel 1996, 75, 931−939. (33) Luo, G.; Yao, H.; Xu, M.; Gupta, R.; Xu, Z. Proc. Combust. Inst. 2011, 33, 2763−2769. (34) Abad-Valle, P.; Lopez-Anton, M. A.; Diaz-Somoano, M.; Juan, R.; Rubio, B.; García, J. R.; Khainakov, S. A.; Martínez-Tarazona, M. R. Fuel. 2011, 90, 2808−2811. (35) Carey, T. R.; Skarupa, R. C.; Hargrove Jr., O. W. Enhanced Control of Mercury and Other HAPs by Innovative Modifications to Wet FGD Processes, Phase I Report for the U.S. Department of Energy, Contract DE-AC22-95PC95260; U.S. Department of Energy: Washington, DC, Aug. 28, 1998. (36) de Visser, E.; Hendriks, C.; Barrio, M.; Mølnvik, M.; de Koeijer, G.; Liljemark, S.; Le Gallo, Y. Int. J. Greenhouse Gas Control 2008, 2, 478−484.

2794


Mercury Release and Speciation in Chemical ... - ACS Publications

Recommend Documents